Momentic offers Mo, an AI QA agent that tests apps end to end

Momentic says Mo tests an app end to end and can find bugs from a prompt. Tests are written as plain-English steps and assertions in YAML files that are stored in the repository and run with the momentic CLI, either locally or in CI. The CLI's exit code can be used as a required check before a merge or deploy.

The company's MCP server lets coding agents start browser sessions, run step ranges and author a test, which is checked in a live session before its YAML is saved alongside the code. A separate mobile MCP server covers iOS and Android emulators from Claude Code or Cursor. Failures are classified as application bugs, test flakiness or environment issues, with screenshots and traces for each step.

When the UI changes, Momentic says it finds controls again during a run and proposes test-file repairs for review instead of applying them directly. A Knowledge base holds product terminology and expected behavior that AI steps consult. Customer figures on the page, such as 117,010 bugs caught, are company-reported.

Teams building agents can have a coding agent write and check end-to-end tests for a change before a pull request merges, using the MCP server or CLI. Proposed test repairs still need review, so a person approves test edits before they land.